Beliefs; The just-war tradition, its last-resort criterion and the debate on an invasion of Iraq.

War is justified only as a last resort; all peaceful alternatives must be exhausted.

That is one of the pillars of the venerable tradition of moral reasoning known as just-war theory, now being cited on every side of the debate about invading Iraq; and however straightforward and sensible the notion of going to war only as a last resort may sound in principle, the world is demonstrating how difficult it is to apply in practice.

To begin with, the problem is both rhetorical -- in the sense of persuasive use of language -- and logical. The criterion of ''last resort'' has been invoked by opponents of virtually every American military action in recent times, from the Persian Gulf to Bosnia, Kosovo and, most recently, Afghanistan; and one has to grant that the opponents can appeal to a kind of logic that makes their opposition unassailable.

In the latest issue of The New York Review of Books, that logic is explained and challenged by Michael Walzer, the political scientist whose book ''Just and Unjust Wars'' has become a classic in the field.

''Lastness,'' he writes, ''is never actually reached in real life: it is always possible to do something else, or to do it again, before doing whatever it is that comes last.'' One more diplomatic initiative, one more peace conference, one more appeal to world opinion, one more nonmilitary form of pressure -- the last resort is always just over the horizon.

This kind of argument is reminiscent of the ancient philosophical conundrum in which Achilles can never overtake a tortoise that has a head start because every time Achilles reaches a point where the tortoise has been, the tortoise will have moved on.

In the real world, of course, Achilles overtakes the tortoise. In the real world, endless diplomatic initiatives can reach a point of diminishing returns, indeed of negative returns. Why would anyone strain the moral criterion of last resort, as a significant number of religious people do, in a way that makes it clear that virtually no military action, short of repelling perhaps a full-scale enemy attack already launched on the nation's territory, would ever meet it?

The reason is not mysterious. A good number of those invoking the just-war criterion of last resort are in reality absolute pacifists opposed to all use of armed force. Or they are what their critics call functional pacifists, not exactly avowing principled pacifism but just never encountering an American use of force they could not denounce. Or they are what might be called isolationist pacifists, for whom nothing except being subjected to that full-scale attack would be irrefutable evidence that the moment of last resort had arrived.

Of course, when pacifists of these various kinds demand the continuation of weapons inspections and surveillance in Iraq as peaceful alternatives to war that are not yet exhausted, they put themselves in a funny position. They know that those inspections are occurring only because of the mobilization of American (and British) military forces. Yet not long ago the same people were opposing that very military mobilization on the basis that all peaceful alternatives had not yet been exhausted.

So for all the talk of going to war only as a last resort, is this criterion really meaningless? Is the principle that war can be morally justified only if all peaceful alternatives have been exhausted a moral standard that can logically never be met?

There is an old Latin maxim in legal and moral reasoning that seems pertinent here: ''abusus non tollit usum,'' abuse does not nullify use. For those like Professor Walzer who value the just-war tradition as a disciplined way to think about the morality of war, the fact that some people have stretched the criterion of last resort to the breaking point in order to support their foregone conclusions does not invalidate it.

The criterion, these just-war theorists say, is essentially a prudential one. Establishing what is or is not a last resort is a matter not of abstract mathematical demonstration but of practical, concrete wisdom, acquired through experience and reflection.

War must be the last resort, Professor Walzer writes, ''because of the unpredictable, unexpected, unintended and unavoidable horrors that it regularly brings.'' As for the notion of lastness, it is essentially ''cautionary,'' he states: ''look hard for alternatives before you 'let loose the dogs of war.' ''

As a matter of fact, when it comes to Iraq, Professor Walzer believes that ''even at this last minute, there still are alternatives, and that is the best argument against going to war.'' The alternatives he outlines focus on containment and control -- maintaining the economic sanctions, the no-flight zones and the United Nations inspections almost indefinitely -- and on fashioning a strong international system that would take responsibility, including military responsibility, for those tasks.

Others, of course, disagree, including both those who would rely more on deterring Iraq's use of arms rather than on preventing their development, and those who have come to see war as the only solution. That argument is being conducted elsewhere, indeed everywhere. What is significant for the just-war tradition is that Professor Walzer himself, despite his admonitions against strained renderings of the last-resort criterion, still considers it central to the current debate.